OAKTrust Digital Repository (Texas A&M Univ)

Kalodimos, Mike

Kalfin, Brian

Wasser, Jim

Achieving Todays Sustainability Goals Utilizing Technology Developed Over Thirty Years Ago: Gas Seals for Process Pumps

Matsutani, Kazumichi

Milne, Reuben

Blackman, Martin

Toci, Emiliano

Major Pipe Support Modification Without Compressor Nozzle Disconnection at LNG Facility

Kurz, Rainer

Wygant, Karl

Brun, Klaus

Process Control for Integrally Geared Compressors

Jahrmarcht, Steffen

Qualification of Additive Manufactured Titanium Impellers for Integrally Geared Compressors to Support Carbon Dioxide Compression

Hardy, Chris

Lungrin, Jeff

First, Rick

Precision Grouting of Turbomachinery Per API 686

Marscher, William D.

Onari, Maki M.

Coppola, Jason D.

An End-User's Guide to Centrifugal Pump Rotordynamics

Racic, Zoran

Yu, John J.

Peton, Nicolas

Deboer, Philip

High Vibration at Motor Barring Speed and its Resolution on a Steam 
Turbine Generator
John J. Yu, PhD
Global Technical Leader
Machinery Diagnostic Services
Bently Nevada
Baker Hughes
USA
Nicolas Péton PhD
Global Director
Machinery Diagnostic Services
Bently Nevada
Baker Hughes
France
Zoran Racic
Lead Machinery Diagnostics Engineer
Machinery Diagnostic Services
Bently Nevada
Baker Hughes
Australia
Philip Deboer
Senior Electrical Engineer
Power Transmission and Maintenance 
APA 
Diamantina PS 
Australia
1
Authors
2
Dr. John Yu, currently based in the USA with Baker Hughes Company, has extensive experience in rotor dynamic 
research and machinery vibration diagnostics. He joined Bently Rotor Dynamics Research Corporation in 1998, 
which later merged into GE and then Baker Hughes. Since 2011, he has served as Global Technical Leader of 
Machinery Diagnostic Services at Bently Nevada. Dr. Yu has authored over 80 technical papers and received the 
2001 John P. Davis Award by ASME IGTI. He holds a PhD in Mechanical Engineering from the University of Alberta 
and is recognized as an ASME Fellow. Additionally, he serves as an Advisory Committee member to the Asia 
Turbomachinery & Pump Symposium.
Nicolas Péton, with a background in machinery diagnostics and extensive global experience, joined GE in 2006. 
Prior to this, he held positions at Alstom Steam Turbine, Cryostar Expander/Compressor, and in the chemical 
industry (PPG, USA), as well as freelancing for startup activities worldwide. Additionally, he has served as a 
mechanical/acoustical research engineer at research institutes (Technion, Haifa and TU Berlin). Currently, Nicolas is 
the Global Director for Machinery Diagnostic Services at GE and a member of the Texas A&M International Pump 
Symposium Advisory Committee. He holds a Diplome d’ingénieur and a PhD from Université de Technologie de 
Compiègne, France.
Zoran Racic is a highly experienced mechanical engineer with a wealth of knowledge spanning over 35 years in the 
field. He began his career at Alstom before transitioning to Bently Nevada in 2016, where he currently serves as a 
Field Service Engineer. Zoran has accumulated over 20 years of experience in machinery vibration diagnostics 
including troubleshooting and multiplane in-situ balancing of gas, steam, and hydro turbogenerators for customers 
in the Australia/ASIA region. He has been certified as Bently Nevada MDS L2 and ISO VCAT 3.
Philip Deboer is a bachelor electrical engineer with over 20 years of experience in power transmission operation 
and maintenance.
Abstract
This case study involves a steam turbine generator where vibration levels surged 
from 20 µm pp to over 200 µm pp during two shutdown events, prompting the 
barring motor to trip due to high vibration. Investigation revealed the cause to be a 
malfunctioning jacking oil pump, leading to rotor-bearings contact during the 90+ 
minutes of barring without support. Repairing the pump restored normal vibration 
levels, emphasizing the importance of obtaining reliable vibration data, even at 
barring speed, for effective diagnosis. Trend, polar, and shaft centerline plots 
proved crucial in identifying malfunctions.
3
1. Introduction
2. Problem Statement
3. Data Review
4. Conclusion
5. Recommendation
6. Resolution and Final Vibration results
7. Lesson learned.
Outline
4
1. INTRODUCTION
The station comprises two blocks with a combined capacity of 242 MW. Each 
block comprises two gas turbine (GT) units and one steam turbine (ST) unit, with 
each unit rated at approximately 40 MW. The steam turbine generator unit 
includes a steam turbine coupled to a gearbox to decrease the turbine speed 
from 5150 rpm to 1500 rpm. The output shaft of the gearbox is connected to a 
4-pole generator.
5
Turbine DE 
bearing #1
Gearbox input shaft 
DE bearing #3
Gearbox output shaft 
DE bearing #4
Barring gear motor
2. PROBLEM STATEMENT
• High vibration issue occurred at 84 rpm during barring speed at turbine bearings as 
well as in the axial direction.
• Barring motor tripped on the motor heat protection, when the shaft rolling was 
subjected to high friction due to rubbing. At that moment, the shaft vibration 
exceeded 200 µm pp (8 mil pp), and the thrust position fluctuated from 0.28 to 0.41 
mm (11 to 16 mil).
• Initial attempts to manually rotate shaft failed; rotor found free next morning.
• Station remained at standstill for further investigations and the OEM inspection, as a 
result from the barring motor fuse issue.
• OEM inspection revealed coupling spacer 5 mm shorter than spec and rubs on 
gearbox DE oil labyrinth seal face. Initial thoughts were that the rub was due to the 
spacer length and high vibrations attributed to rotor bending, possibly from steam 
turbine ingress.
6
Rundown 
from 5150 
rpm to 
barring 
speed 84 
rpm for 26 
minutes
Vibration low with thrust bearing 
position steady for the initial 62 
minutes 
Vibration increased to over 200 µm
pp (8 mil pp) when the shaft
jammed, and barring motor tripped
due to overheating protection.
At motor barring speed of 84 rpm for 114 minutes
200 um pp 
(8 mil pp)
100 um pp 
(4 mil pp)
150 um pp 
(6 mil pp)
50 um pp 
(2 mil pp)
3.1 Data Review-DCS readings for the 1st trip event
7
• After the OEM inspection, and before the run up to rated speed of 5150 rpm, 
the unit was on barring  speed of 84 rpm for 141 minutes, 27 minutes longer 
than during the 1st trip event.  Vibration was low and stable at both, barring 
and rated speed.
• Further inspection and remedy work of fitting shims to adjust for the quill shaft 
length were scheduled for the next outage.
3.2 Data Review- for the 1st trip event
8
• Identical vibration excursions observed at the barring speed (84 rpm) during another 
data collection session.
• Vibration remained steady and low for the first 65 minutes.
• Afterwards, vibrations started trending up slowly for 20 minutes.
• In the next 30 minutes, vibration increased abruptly, reaching 164 µm pp (6.5 mil pp).
• Decision made to take the unit off barring to prevent further bearing/seal damages.
3.3 Data Review- for the 2nd trip event
9
Barring speed 84 rpm
5150 rpm 
Shutdown Vibration spikes
3.4 Data Review- Brg#1 direct & 1X vibration leading to  the 2nd trip
Vibration spikes
Barring speed 84 rpm
5150 rpm 
Shutdown
Brg#1X vibration Brg#1Y vibration
10
Jacking oil pump OFF Jacking oil pump OFF
3.4 Data Review- Thrust bearing vibration and position on the 2nd trip
Barring speed 84 rpm
5150 rpm 
Shutdown
Barring speed 84 rpm
5150 rpm 
Shutdown
Thrust probe vibration Thrust probe gap
11
Jacking oil pump OFF Jacking oil pump OFF
• Multiple abnormal shifts in gap voltage from radial proximity probes occurred.
• Vibration spikes (not-1X) also accompanied by changes in gap, likely due to sudden 
changes in gap voltage.   
3.5 Data Review- Brg #1 vibration vs. gap on the 2nd trip
Jacking oil pump issue? 
Brg #1X Gap Brg #1Y Gap 
Brg #1X Brg #1Y
12
Jacking oil pump ON
Jacking oil pump OFF
Jacking oil pump ON
Jacking oil pump OFF
In normal conditions, 
the DC gap shouldn't 
exhibit sudden 
frequency shifts.
• Sudden journal position changes due to the jacking oil pump on/off events.
• Shifts in gap trend indicate these events.
The shaft 
position change 
at 84 rpm due 
to probes 
thermal 
sensitivity.
At 84 rpm shaft 
dropped for 37 
µm (1.5 mil) 
when the jacking 
oil pump turned 
off (tripped).
3.5 Data Review- Brg #1 shaft centerline/gap on the 2nd trip
Brg #1 
Brg #1X Gap Brg #1Y Gap 
13
At 84 rpm shaft 
dropped for 60 
µm (2.4 mil) 
when the jacking 
oil pump turned 
off (tripped).
Turbine bearing #2 – shaft centreline and DC Voltage gap trends
3.6 Data Review- Brg #2 shaft centerline/gap on the 2nd trip
• Sudden journal position changes due to the jacking oil pump on/off events.
• Shifts in gap trend indicate these events.
Brg #2 
Brg #2X Gap Brg #2Y Gap 
14
1X vector rolling (against shaft rotation) and increasing
3.7 Data Review- Polar plots at Brg #1 on the 2nd trip
Brg 1X Brg 1Y
15
Abnormal excursions in 
the 1X vector were 
indicative of shaft 
thermal bow resulting 
from rubbing caused by 
contact between the 
shaft and stationary 
parts.
3.8 Data Review- Polar plots at Brg #2 and thrust probe on the 2nd trip
• The 1X vector observed at the radial 
probes demonstrates a radial outward 
movement, likely attributed to the 
low-speed conditions where the hot 
spot aligns with the high spot when a 
rub occurs, or its continuous bow with 
the same direction.
• The 1X vector at the thrust probe 
displays a similar vibration pattern, 
possibly influenced by the significant 
shaft bow, which impacts the 
oscillation of the cross-section area 
monitored by the thrust probe.
Thrust probe
Brg 2X Brg 2Y
16
Under normal circumstances, 
polar plots shouldn't show 1X 
vibration vectors increasing 
from ~0 to 160 µm pp (6+ mil pp) 
at the barring speed of 84 rpm.
 
• A larger orbit was observed at Brg#2 (outboard).
• The shaft centerline position appeared lower at Brg#1 (inboard), possibly due to its 
weight center location.
• It's worth noting that the shaft centerline plots may not accurately reflect the true 
position relative to the bearing due to the thermal effect of proximity probes and 
casing.
3.9 Data Review- DC orbit overlay at Brg#1 and #2 on the 2nd trip
Brg#1 Brg#2
17
• Starting at around 300 rpm during the coastdown, the jacking oil pump must be turned 
on. 
• However, the jacking oil pump was ON and OFF service for multiple times.
• What was a root cause for the jacking oil pump frequent tripping – getting OFF the 
service?
  Two possibilities: 
➢ The jacking oil line pressure being too high due to oil flow restriction such as cloggage, 
the pump motor pulling too much current and thus overheating to lead to a 
protection trip.
➢ The motor having electrical problems, such as loose phases terminals or fuses.
3.10 Data Review- Analysis
18
• The high vibration excursions experienced at the barring speed (84 rpm) in two 
separate events were attributed to the operation issue with the jacking oil 
pump.
• At the barring speed of 84 RPM, the deteriorated jack oil pump/system did not 
lift the shaft adequately, leading to rubbing occurrences.
• The insufficient shaft lift resulted in shaft thermal bow, intensifying rubbing, and 
eventually causing possible rubs between the shaft and turbine bearings.
4. Conclusions
19
• Identify the unusual operation of the jacking oil pump, and inspect the pump motor, 
pump system, filters, and pipes.
• Inspection revealed a phase terminal issue on the pump motor, prompting the station 
to opt for replacement rather than attempting repairs.
• The station decided to replace both the pump and motor with new ones from their 
spare storage to address the issue effectively.
5. Recommendations and Actions 
20
• Following the replacement of the jacking oil pump and motor within a week, the 
unit was scheduled to return to service.
• A testing procedure was set, requiring the unit to remain at the barring speed for at 
least 3 hours during ramp-up and coast-down.
• The station extended the barring period significantly, keeping the unit on barring for 
nearly 15 hours before startup.
• After synchronization and loading for 5.5 hours, the unit underwent a hot rundown, 
followed by more than two hours at the barring speed.
• Throughout the prolonged barring period, vibrations remained consistently low and 
steady, maintaining historical levels across the steam turbine generator train 
bearings.
• Despite the forced outage, the turbine-to-gearbox coupling spacer was not 
replaced.
6.Resolution and Final Vibration results
21
84 
rpm
barring speed 84 rpm 84 
rpm
Barring speed 84 rpm
84 
rpm
barring speed 84 rpm 84 
rpm
Barring speed 84 rpm
6.Resolution and Final Vibration results (Cont.)
Brg #1X Brg #1Y
Brg #2X Brg #2Y
Vibration low and stable at barring speed of 84 rpm 
after replacing with the new jacking pump and motor    
22
• Maintaining the jacking oil pump in optimal condition is vital to elevate the rotor, 
avoiding it from rubbing against stationary components during barring speed. 
Such rubbing could lead to significant vibration spikes, especially following a hot 
shutdown.
• Acquiring vibration data, even during barring speed, is essential for effectively 
diagnosing machine malfunctions. Without reliable vibration data, identifying the 
true root cause of issues becomes challenging. Trend, polar, and shaft centerline 
plots serve as highly valuable tools for detecting machine malfunctions, even at 
barring speed.
• A quick resolution is imperative to prevent catastrophic machine damage. Failure 
to act promptly could result in severe rubbing and permanent shaft bowing.
7. Lessons learned
23
• There are other surveillance parameters that could have contributed to 
satisfactory warning and resolution, such as motor amps, jack-oil pressures, etc. 
• A Fishbone diagram (or Ishikawa diagram) could prove to be highly beneficial for 
systematic problem-solving, particularly in cases like this.
7. Lessons learned (Cont.)
24
Thanks….. and Questions (?)
25


High Vibration at Motor Barring Speed and Its Resolution on a Steam Turbine Generator

Simon, Marco

Gelin, Alain

Schindler, Peter

Bellocq, Pablo

Controlled
RESOLUTION OF CENTRIFUGAL COMPRESSOR FIRST MODE
INSTABILITY OBSERVED DURING FULL SPEED FULL LOAD STRING TEST
Marco Simon
Peter Schindler
Alain Gelin
Pablo Bellocq
Controlled
Marco Simon is the heading the Product Integrity – Europe within Compression at
Siemens Energy. He is responsible to drive global troubleshooting, technical
support and complex Root Cause Analysis. Before, he was responsible for fact
finding, repair, troubleshooting and revamps in Service Technical Support.
Peter Schindler is a troubleshooting engineer working for Siemens Energy in
Duisburg, Germany. His main expertise in rotating machinery is blade vibrations
and vibration analysis.
Author Biographies
Dr. Pablo Bellocq, Rotating Equipment Specialist at TotalEnergies. He graduated as
mechanical engineer and PhD in power and propulsion. Before joining TotalEnergies
he worked for Safran Helicopter Engines, CAF (railway systems) and UTE (Ugruguayan
power gen/distribution corp.)
Dr. Alain Gelin, Rotating Equipment Expert at TotalEnergies. He graduated as
mechanical engineer and PhD in rotodynamics. Before joining TotalEnergies he
worked for Thermodyn as R&D and testing technical leader.
Controlled
Abstract
High sub-synchronous vibrations were observed in the LP/IP section of a
25 MW, 2 casing turbo compressor during a full speed/full load string
test. The vibrations were repetitively observed at the natural frequency
of the first lateral rotor mode of the LP/IP compressor when reaching a
certain gas density.
Additionally, a torsional natural frequency was observed at this exact
same frequency, with the highest amplitude at the coupling connecting
the LP/IP compressor to the gearbox.
This case study presents the observations, analysis and succession of
testing done to understand and correct the instability issue.
Controlled
Compressor Details (LP/IP)
• Train configuration:
Gas turbine – LP/IP compressor
– Gear box – HP compressor
• Process gas: Natural gas
• Inlet pressure: 3.36 bar(a)
• Inlet temperature: 24.9°C
• Discharge pressure: 41.01 bar(a)
• Mass flow (design): ~150,000 kg/h
• Total power: 25 MW
• Compressor design:
Back-to-back arrangement,
5 (LP) + 4 (IP) stages
• Maximum continuous speed: 6825 rpm
GT
LP/IP Gear
HP
Controlled
Observations During String Test
• Sub-synchronous vibrations at the
LP/IP compressor occur exactly at
the natural frequency of the 1st
lateral rotor mode.
• Repeatedly, when reaching ≈1/3 of
rated gas density, sub synchronous
vibrations reach trip values
(instability).
• A torsional natural frequency was
observed at the same frequency
with the highest amplitude at the
coupling connecting the LP/IP
compressor to the gear box.
Sub-sync. Vibration/
Instability
1X Rotor Speed
Frequency
Controlled
Initial Investigation Steps
• Objective was to understand the cause of the lack of system damping.
• Initial reviews were carried out: design, manufacturing and test
records.
• The following hypotheses (H) were considered:
H1: Overestimation of system damping
H2: Major discrepancy between design and manufacturing
H3: Destabilization caused by non linearities in disk coupling on a
shaft-line running with first lateral mode of LP/IP compressor in
coincidence with second torsional critical speed.
Controlled
• Instability was observed at only 1/3 of gas
density corresponding to the predicted
instability.
• Verification of:
• cross-coupling forces by reviewing all
calculations
• bearing characteristics
• lateral stability analysis (internal and
3rd party)
• No finding justifying instability.
 Hypothesis 1 was ruled out.
Affected compressor
Cross-Coupling
Lo
g.
 D
ec
.
1.0
0.4
Max Clr
Min Clr
Cross-Coupling at
Design Conditions
H1: Overestimation of System Damping
-0.2
Av. Gas Density
Fl
ex
ib
ili
ty
 R
at
io
0.1
Controlled
• All manufacturing records were reviewed.
• Bundle was opened and inspected:
• General inspection (assembly checks)
• Dimensional and geometrical checks of rotor and stator
• Bearing checks (visual & full micro geometry measurements:
clearance, pad curvature => preload)
• Labyrinths
• Rubbing
• No findings.
 Hypothesis 2 was ruled out as main root cause.
H2: Manufacturing/Assembly Issue
Controlled
H3: Destabilization from the Coupling (1)
• Coincidence of 1st lateral mode of LP/IP with 2nd train torsional mode
(indicated in red) with large amplitude at LP/IP shaft ends (DE/NDE):
• Destabilizing cross coupling forces could come from non-linearities and
internal friction in the coupling disk packs under torsional oscillation
magnified by the coincidence.
 Decision: Disconnect LP/IP-Gearbox and run uncoupled
Driver (free power turbine of GT) LP/IP
Gear box
HP
Couplings
Controlled
• Operational modal analysis (OMA) was used for determination of
system damping during compressor solo run.
• Instability also observed and LogDec was significantly lower than
predicted even with very low gas density.
H3: Destabilization from the Coupling (2)
Logarithmic DecrementTest Point
0.06OMA #1 (10% Test Load)
0.06OMA #2 (50% Test Load)
0.01OMA #3 (100% Test Load)
 Conclusion: Instability is neither originating from the coupling/
a coupled lateral-torsional phenomenon nor from aerodynamic
effects.
Controlled
Compressor Solo Run: Observations
• At critical speed, system damping is as expected (Log. Dec. ~0.5)
• At full speed, system damping is much lower than expected (measured
~0.03 – 0.07; predicted ~0.3)
Max continuous speed
Max clearance
Min clearance
Speed [rpm]
Lo
g 
D
ec
 1
st
fw
d
m
od
e
Speed [rpm]
Am
pl
itu
de
 [µ
m
]
40003000200010000
AF = 6.4
Log. Dec. = 0.5
Controlled
Compressor Solo Run: Conclusions
• At this point we knew that something independent from gas
destabilizing cross coupling forces was causing the “degradation” of
damping between first critical speed and 100% speed.
• H4: Internal friction within the rotor reduces system damping.
• A design and rotor assembly review was done and concluded that even at full
speed, none of the shrink fits were loose (which could have led to rotor internal
friction).
 Hypothesis 4 was ruled out.
• H5: Loss of damping due to destabilizing cross coupling forces
generated by friction inside the bearing.
• Bearings are of tilting-pad type with spherical support (ball-and-socket design).
• There is no straight forward approach of estimating  friction and resulting
magnitude of destabilizing cross coupling forces.
 A bearing inspection was done after the solo run.
Controlled
H5: Loss of Damping due to Bearing Pad Friction
After only a few hours of operation, very clear rubbing
marks were observed (including fretting) on the balls of
the ball-and-socket support, and at the back of the pads.
Controlled
H5: Loss of damping due to bearing pad friction
It was decided to perform the following test runs including OMA:
1. Open air run at full speed with original job bearings (ball-and-socket type)
To confirm that OMA can verify a low LogDec (0.06-0.01) even when running at open
air (marginal gas destabilizing effects).
2. Open air test with double curvature pad (rockerback) bearings
Bearings having comparable definition (L/D, clearance, preload, offset).
To confirm by OMA that LogDec was “as predicted” (~0.3) at full speed, which would
confirm that the friction at the back of the pads is the cause of loss of system damping.
Test 1 Test 2
Controlled
H5: Loss of Damping due to Bearing Pad Friction
Logarithmic Decrement
Condition
Double CurvatureBall & Socket
0.310.07Tested @100% speed
0.190.03Tested @105% speed
0.13 – 0.260.16 – 0.28Predicted @105% speed
• Damping with double curvature bearings is significantly higher than with the
specific ball-and-socket type bearing originally built for this application.
• Damping w/ double curvature bearings is in good agreement with prediction.
Open air test: OMA-Results compared with prediction:
Controlled
Corrective Actions
• Job bearings (ball-and-socket type) were replaced with double
curvature bearings.
• To further increase system damping, it was decided to install squeeze
film dampers and a hole pattern seal at the division wall.
• Couplings were changed to have enough separation margin between
torsional natural frequencies, running frequencies and lateral
frequencies of the shaft line.
Controlled
Conclusion – Lessons Learned
• Full load string tests are very valuable to validate the design of
turbomachinery packages before delivery from the factory.
• Friction at the back of tilting pads can create significant destabilizing
cross-coupling forces (known from theory).
• These cross-coupling forces are usually not considered in design stage as they are
difficult to estimate, not significant if the right geometry and material is selected.
• This should be considered when selecting and designing ball-and-socket pad type
bearings.
• Operational modal analysis (OMA) is a useful tool for evaluation of
system damping in compressor tests. OMA can be used for
troubleshooting and even for preliminary validation of rotodynamic
predictions during standard mechanical run tests or PTC-Type 2 tests.


Resolution of Centrifugal Compressor First Mode Instability Observed During Full Speed Full Load String Test

BKAM, Full DALA (Database of African American and Predominantly White American Literature Anthologies) Corpus, Morrison, Author Page Count, as a percentage, zoom.Referenced in Chapter 4 of the book " Digital Literary Redlining: African American Anthologies, Digital Humanities, and the Canon.

Texas A&M University

Not a member yet

Achieving Todays Sustainability Goals Utilizing Technology Developed Over Thirty Years Ago: Gas Seals for Process Pumps

Major Pipe Support Modification Without Compressor Nozzle Disconnection at LNG Facility

Process Control for Integrally Geared Compressors

Qualification of Additive Manufactured Titanium Impellers for Integrally Geared Compressors to Support Carbon Dioxide Compression

Precision Grouting of Turbomachinery Per API 686

An End-User's Guide to Centrifugal Pump Rotordynamics

High Vibration at Motor Barring Speed and Its Resolution on a Steam Turbine Generator

Resolution of Centrifugal Compressor First Mode Instability Observed During Full Speed Full Load String Test

Black Fire authors

What's in My Water?

47,493

136,879